continuous condition monitoring of pipelines and risers ...€¦ · sintef materials and chemistry...
TRANSCRIPT
-
SINTEF Materials and Chemistry 1
Ole Øystein Knudsen, SINTEF
www.smartpipe.com
Continuous condition monitoring of pipelines and risers
-
SINTEF Materials and Chemistry 2
• Technical condition• Warnings• Simulations• Visualisation
Data interpretation
• Materials degradation• Analysis tools• Database
• Sensors• Communication• Power
Decision making
Data collection
-
SINTEF Materials and Chemistry 3
General system requirements
• No interference with laying operations• Non intrusive sensors• Low cost, simple, robust• Lifetime >20 years• Interfacing to existing sensor technology• Low power consumption• Local processing to reduce needed communication capacity
-
SINTEF Materials and Chemistry 4
Lithiumbatteries
-
SINTEF Materials and Chemistry 5
SmartPipe
Sensors• Ultrasound wall thickness
measurement• Strain gauges for measuring
deflection, vibration and internal pressure
• Thermistor for temperature measurements
• Accelerometer for measuring inclination
Communication• Wireless electromagnetic signal in
coating
Power• A package of conventional Lithium
batteries• Thermoelectric generator
-
SINTEF Materials and Chemistry 6
Battery packing
• Four cells in series inserted in steel tubes– 600 mm length– Space available for charge
balancing electronics– Hermetically sealed to prevent
water intrusion
• Ten 14.4 V battery packs encased in exterior bracelet– 40 cells in total (120% capacity)– Close contact to cold sea water
beneficial for battery lifetime
• Exterior bracelet can be partially embedded in PP‐insulation.
-
SINTEF Materials and Chemistry 7
Communication
• Electromagnetic wave in the coating
• Inter node distance: 24 m• Redundancy distance: >200 m (8‐
10 nodes)• Carrier frequency: 3.8 MHz• Output power +14 dBm• Band with: 150 kHz
• Patented
Antenna
Signal
Multihop topography
variable hop lengthintelligent end nodesdead nodes
variable hop lengthsimple nodes intelligent end nodes
-
SINTEF Materials and Chemistry 8
SmartPipe data management
Fibre modem or powerline modem
Host machineData
collecion system
Fibre modem or powerline modem
Cab
le o
r opt
ical
fibr
e
Data base
Sub-sea
Fibre modem or powerline modem
Pick-up antenna on outside of pipe
Cable or optical fibre
Degrada‐tion analysis
Visuali‐sation tools
-
SINTEF Materials and Chemistry 9
Test pipe: 216 m long 10" pipe 80 mm PP coating8 sensor and communication nodes
Testing communication range and sensorsInstalled November 18, 2013 in the Orkanger fjord on 10‐15 m water depth
SmartPipe Meso test
-
SINTEF Materials and Chemistry 10
• Steel thickness = 9/16 i.e. 14.3mm• OD = 10 3/4• ID = 9 5/8
The SmartPipe
-
SINTEF Materials and Chemistry 11
-
SINTEF Materials and Chemistry 12
Antenna
-
SINTEF Materials and Chemistry
• Communication at least 200 m along the pipe• Able to communicate with 8 nodes along the pipe• Attenuation of the communication signal is as expected or better
13
Communication variable hop length
intelligent end nodesdead nodesvariable hop length
simple nodes intelligent end nodes
-
SINTEF Materials and Chemistry
• Pipe set in vibrational motion on land before the installation
• Counting vibration half cycles in different strain ranges
• Strain gauges mounted around the pipe as shown
• 93 half cycles • Ax2 and Ax4 registered 92 and 93 half
cycles
14
Fatigue monitoring test
0
10
20
30
40
50
60
70
80
90
100
0 200 400 600 800 1000 1200Ra
infall ha
lf cycle coun
tsAmplitude peak to peak (µ‐strain)
0‐Hoop
1‐Axial
2‐Axial
3‐Axial
4‐Axial
5‐Hoop
Ax1
Ax2
Ax3
Ax4
H2
H1
Accelerometer
-
SINTEF Materials and Chemistry 15
Temperature (SINTEF data log)
Holes in data logged at SINTEF are caused by faulty 3G router and PC outages, not by faults in the SmartPipe system.
The figure shows data from two of the four sensors at node 117
All functional nodes and sensors show practically the same temperature with variation in the least significant bit only. This is as expected.
Variation in temperature is probably cased by changes in water temperature due to tides, and weather conditions.
Sensor 2 from last week only
Sensor 1
degrees C
-
SINTEF Materials and Chemistry
Strain (SINTEF data log)The figure shows two of the six sensors at node 119
• The strain values correlate with the temperature. This is as expected.• As expected the strain values do not change much over time (just a few μS)• The static strain value on each sensor element is more or less random and assumed to be caused by the
mounting procedure (stretching of belt etc)
• Some sensors show some outliers that have been removed from these plots.
microstrain
Strain
Temperature
linear regression curve
-
SINTEF Materials and Chemistry 17
Ultrasound thickness (SINTEF data log)The figure shows two of the eight sensors at node 119
rising = more diffuse echo
echo quality
thickness
unit 10µm13.080 mm
• For most sensors the data shows a falling thickness and falling echo quality (rising quality number).• The mean change of thickness corresponds to 100um/year which is within expected range (steel in ocean water).• The echo quality is expected to fall as the steel surface corrodes.• Some sensors are more noisy than others but the averaged value seems to be correct.• Some sensors show outliers that have been removed from these plots.
12.475 mm
linear regression curve
-
SINTEF Materials and Chemistry
• Strain gauges• Ultrasound
18
Laboratory testing of sensor resolution
-
SINTEF Materials and Chemistry
• Tensile testing machine• Compressive mode• 10 to 250 kN• Calculated strain plotted vs measured
strain• 5% scaling deviation• Resolution better than 10 µ‐strain
19
Strain test
-
SINTEF Materials and Chemistry
• Sensor belt mounted on steel panel
• Grinded on the backside in ca 10 µm steps
• Sensor measurement vs µm screw gauge
• 180 µm grinded away• Average of 4 measurements has a
resolution better than 10 µm
20
Wall thickness
-
SINTEF Materials and Chemistry 21
The consortium